Study design

This descriptive-analytical study aimed to investigate the factors influencing participation in CRC screening using data from Iran’s Integrated Electronic Health System (SIB, a Persian acronym meaning “apple”). Launched by the Ministry of Health in February 2015, SIB was designed to modernize health data management. The system currently hosts electronic health records (EHRs) for over 73 million individuals and is operational in more than 36,000 urban and rural areas. SIB is distinguished as the most comprehensive and advanced system, offering online functionality and nationwide integration [39].

Study population

Population in the study had to meet the following inclusion criteria: individuals aged 50–70 years residing in Isfahan province from 2016 to 2021, with information recorded in the SIB system (access date: 5/23/2023). The exclusion criteria were: high-risk individuals for CRC (personal history of cancer, personal history of polyps, IBD, first-degree family history of cancer, second-degree family history of cancer, weight loss, constipation, and bleeding), individuals diagnosed with CRC before the study period, those with no health service records after turning 50, and dead individuals. The study protocol was approved by the ethics committee of Isfahan University of Medical Sciences (IR.MUI.RESEARCH.REC.1398.257). It utilized anonymized secondary data on CRC screening, with all identifying information removed to ensure participant confidentiality. As a result, the need for informed consent was waived by the ethics committee.

Data collection

We developed a form based on Andersen’s Behavioral Model to address predisposing, enabling, and need-for-care factors, as well as the utilization of CRC screening services, using existing variables. Predisposing factors included age, gender, marital status, and education level. Enabling factors included job, health insurance coverage, and nationality. The need-for-care factors comprised comorbidities and lifestyle factors. Comorbidities included hypertension, hyperlipidemia, diabetes, and cardiovascular diseases. Lifestyle factors included physical activity level, obesity, and consumption of salt, oil, vegetables, fruits, and fast food. In this study, the lifestyle variable was used to categorize individuals into three groups: those with the highest scores were classified as the “healthiest group,” those with the lowest scores as the “high-risk group,” and those with intermediate scores as the “low-risk group.”

Statistical analysis

The analysis utilized a logistic regression model to identify factors associated with participation in colorectal cancer (CRC) screening. The likelihood of participation was estimated using a logit model, taking into account predisposing, enabling, and need factors. Data were analyzed using SAS JMP Pro 17. Variables with p < 0.001 were selected as candidate independent variables for multivariable logistic regression analysis. The goodness-of-fit for the multivariable logistic regression models was assessed using the Whole Model Test, Wald test, and Lack of Fit test. Results were presented as odds ratios (OR) with their corresponding 95% confidence intervals (CI) at a significance level of 0.001. Additionally, probabilities were presented based on the model, and predictor screening was performed using 100 bootstrapping iterations. Predictor Screening is a method used to evaluate and rank predictors based on their ability to explain the variability in the outcome variable, which in this case is participation. The contribution represents the total impact of the predictor on the model’s ability to predict participation. A higher value indicates a stronger influence on the prediction accuracy. Portion is the proportion of the total contribution made by each predictor. This gives a sense of the relative importance of each predictor in the model. Predictors are ranked based on their contribution, with 1 being the most significant predictor.

Individuals’ characteristics

A summary of the characteristics of 808,978 average-risk individuals in Isfahan, with a mean age of 56.67 years (SD = 7.17), is presented in Table 1. Over half of the individuals were men, most had elementary education, 87.7% were married, and 97.72% were Iranian. More than half do not have the disease and less than half have a healthy lifestyle. More details are presented in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Average Risk Individuals’ characteristics (N = 808978).

https://doi.org/10.1371/journal.pone.0323291.t001

Over the five years of the screening program, approximately 40% of individuals participated in colorectal cancer screening.

Logit regression

The logistic regression analysis examined the factors influencing participation. The odds ratios (ORs) for changes in the levels of independent variables are presented in Table 2:

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Odds Ratios for changes in the levels of independent variables.

https://doi.org/10.1371/journal.pone.0323291.t002

Predisposing factors

The study considered several predisposing factors, including age, gender, marital status, and education level. Age also had a notable effect, with individuals over 65 having significantly lower odds of participation compared to those younger than 65 (Odds Ratio = 0.388, p < 0.0001). Gender also played a role, as males had higher odds of participation compared to females (Odds Ratio = 0.896 for females, p < 0.0001). Marital status also affected participation, as being part of a couple increased the odds of participation compared to being single (OR = 1.158, p < 0.0001). Education had a slightly negative effect on participation, with a per-unit increase in education reducing the odds of participation (Odds Ratio = 0.919, p < 0.0001).

Enabling factors

Enabling factors, such as employment status, health insurance coverage, and nationality, were also significant. Work Time Pattern also influenced participation, with those having flexible or freelance times being more likely to participate compared to those with fixed or flexible times (fixed vs. free: Odds Ratio = 0.784, p = 0.0001; flexible vs. free: Odds Ratio = 0.813, p < 0.0001). Health insurance coverage was another key enabling factor. Individuals without health insurance were much less likely to participate compared to those with insurance, while individuals with insurance were almost twice as likely to participate (Odds Ratio = 1.852, p < 0.0001). Non-Iranians had significantly lower odds of participating compared to Iranians (Odds Ratio = 0.694, p < 0.0001).

Need factors

Need factors, which include the presence of chronic diseases and lifestyle factors, were among the most influential. The presence of chronic diseases increased the odds of participation, with a gradient effect observed: One chronic disease (OR = 1.972, p < 0.0001), two chronic diseases (OR = 2.521, p < 0.0001), and more than three chronic diseases (OR = 2.983, p < 0.0001). Individuals with high-risk lifestyles were significantly more likely to participate compared to those with healthy lifestyles (OR = 1.849, p < 0.0001) but low-risk lifestyle participants had similar odds to healthy individuals (OR = 0.991, p = 0.2433).

Model fit

The model demonstrated a good fit based on the Whole Model Test. Key fit statistics included a -LogLikelihood Difference of 34,158.13 and a Chi-Square value of 68,316.27 with a p-value less than 0.0001. The results of the lack of fit test show a Chi-Square value of 11,088.52 and a p-value less than 0.0001, indicating a significant lack of fit. Effect Wald Tests for participation indicate several significant predictors. The key effects, ranked by their significance (Logworth), include disease, education, age, lifestyle, insurance, job type, marital status, nationality, and sex, all with a p-value of 0.00000, demonstrating highly significant impacts on participation.

Probability

Gender differences are also observed, as females exhibit a lower probability of participation compared to males. The probability of Participation is greater for the Age Group=>65 than <65. Marital status impacts participation, with a couple of individuals having a slightly higher probability of participation compared to those in singles. Furthermore, insurance coverage positively influences participation; individuals without insurance have a lower probability of participation, whereas having insurance increases the probability. Individuals with fixed or flexible job types have a lower participation probability, while those with free job types show higher participation probabilities. Education level also affects participation, with a slight but noticeable decline in participation probability as education level increases. This suggests that individuals with higher education levels tend to participate less. Nationality is another influential factor, with Iranians showing a higher participation probability compared to non-Iranians. Lifestyle is a significant determinant as well; healthy individuals have the lowest participation probability, which increases for individuals with low-risk lifestyles and further increases for those with high-risk lifestyles.

Fig 1, visually represents the effects of various predictors on participation probabilities from the logistic regression analysis. This provides a comprehensive overview of how each factor affects participation, aligning with the logistic regression analysis results. Understanding these effects can guide policy-making and interventions aimed at improving participation rates. This visual summary presents common scenarios, including the best and worst-case scenarios, with probabilities of participation at 3% and 90%, respectively.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Prediction Profiler: Effects of Predictors on CRC Screening Participation Probabilities.

A: Common Scenario – Displays actual predictor values and their associated participation probability. B: Worst-case Scenario – Demonstrates the lowest probability of participation (3%) under unfavorable predictor conditions. C: Best-case Scenario – Shows the highest probability of participation (90%) under optimal predictor conditions.

https://doi.org/10.1371/journal.pone.0323291.g001

Predict participation

Table 3 summarizes the contribution and importance of each predictor in determining participation. The predictor screening analysis reveals that lifestyle, chronic disease, and education are the top three factors influencing participation, with lifestyle being the most dominant predictor. This information can be used to target interventions and policies more effectively, prioritizing changes in lifestyle and health management to enhance participation rates.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Relative Contribution and Ranking of Predictors Influencing Participation.

https://doi.org/10.1371/journal.pone.0323291.t003

Gender:

Our study found that males had a higher probability of participating in CRC screening than females with other variables in the model. This is in line with findings from studies that reported higher CRC screening rates among men compared to women[10–12]. The reasons for this may be linked to women’s reported discomfort or fear about the screening process [12]. However, some studies show opposite results, found no significant gender differences in CRC screening uptake [13], or found Overall screening uptake is higher in women than men [14]. Thus, barriers to CRC screening may differ by sex [15], This discrepancy may be due to cultural or healthcare system differences, suggesting that gender-related participation might be more context-specific.

Age:

Our study confirmed that younger adults had higher participation rates, aligning with the findings of Koo et al. (2012) in the Asia-Pacific region. Koo’s study indicated that respondents aged 50 and above from Hong Kong, Indonesia, Korea, and Taiwan exhibited lower awareness of symptoms, risk factors, and, to a lesser extent, screening tests compared to their younger counterparts. However, no significant age-related differences were observed in the other countries studied [16]. Chan et al. (2022), observed that younger individuals were more likely to participate in colorectal cancer screening [13]. However In contrast, Huang et al. (2021) reported that older individuals (aged 66–70) were more likely to participate, potentially due to targeted awareness campaigns within those populations [17]. This highlights the variability in age-related participation rates depending on public health strategies and underscores the need for interventions aimed at older age groups.

Marital status:

Marital status Marital status plays a significant role in influencing participation in health screenings. Research by Gram et al. (2020) found a strong link between marital status and CRC screening, with participation rates being particularly high among married individuals, especially those whose partners also participated [18]. Similarly, El-Haddad et al. (2015) reported that in the USA, both married individuals and those in unmarried partnerships had higher odds of undergoing CRC screening compared to other marital status groups[19]. Chan et al. (2022) observed comparable trends in multi-ethnic Asian populations, where married people were more likely to engage in health services, including CRC screening [13].

Education:

Interestingly, in our study, higher education was linked to lower participation in screening (OR = 0.920), which challenges the common belief that greater education enhances health literacy and preventive care behaviors. Willems and Bracke (2018) found that individuals with higher education levels were more likely to participate in screenings for cervical, breast, and colorectal cancer compared to those with lower education levels. However, they also noted that educational disparities were smaller in countries with organized cervical cancer screening programs than in those with opportunistic screening[20]. Budhraja et al. (2011) highlighted that patient education is a powerful but underutilized tool for improving the effectiveness of CRC screening[21]. These contrasting results suggest that while education typically boosts health literacy, its impact on screening participation may vary based on factors such as trust in healthcare systems and access to screening services. Our findings indicate that individuals with higher education levels participated less in the screening program. The mean years of education in the study population was approximately 8 years (SD = 4.89), suggesting a relatively low overall education level. However, since the screening program was conducted within the healthcare system and primarily targeted rural populations, participation was naturally higher in rural areas compared to urban regions. Additionally, as these services were provided during standard working hours, individuals with higher education—who are more likely to be employed in fixed-hour jobs—may have faced more barriers to participation. Another possible explanation is that trust in the healthcare system may be lower among more educated individuals, influencing their willingness to engage in such programs. Given these factors, the observed results are not unexpected.

Nationality:

Our results indicated that Iranian nationality was positively associated with higher participation in colorectal cancer (CRC) screening (OR = 1.442). Warren Andersen et al. (2019) found lower CRC screening rates among African American individuals and those with low socioeconomic status[22]. Similarly, a study in Norway revealed that immigrants, particularly from non-Western countries, participated less frequently in a CRC screening pilot compared to non-immigrants [23]. This lower participation among immigrants may be attributed to factors such as language barriers, cultural beliefs, and limited awareness about screening programs.

Work time pattern:

Employment type influenced CRC screening participation in our study, with individuals in flexible or freelance work time patterns showing higher participation rates. Similarly, Fedewa et al. (2022) reported that unemployment was linked to lower cancer screening rates, largely due to a lack of health insurance[24]. Collatuzzo et al. (2022) highlighted that cancer disparities are partly rooted in occupational factors such as job type, position, tasks, work schedule, salary, and employment status. These factors contribute to varying cancer screening participation across different working categories, reflecting disparities in access to health services[25].

Insurance:

Individuals with health insurance are more likely to pursue preventive care, including CRC screening, among U.S. adults[26]. Gawron et al. (2021) noted that the lack of insurance coverage and the high cost of care were significant barriers to CRC screening[27]. However, in a study by Jones et al. (2022), health insurance status was not found to be associated with the earlier stage of diagnosis or the mode of detection in a diverse group of patients recently diagnosed with colon cancer[28]. This suggests that while insurance influences screening participation, it may not necessarily impact the stage at which cancer is detected.

Chronvic diseases:

Our study found that individuals with chronic diseases were more likely to participate in CRC screening. Coronado et al. (2022) observed an inverted U-shaped relationship between a patient’s chronic disease burden and a provider’s recommendation for a FIT and a negative linear relationship between chronic disease burden and FIT completion[29]. Similarly, Guiriguet et al. (2017) found that having three or more major chronic conditions was associated with lower participation in FIT-based CRC screening programs, while individuals with several minor chronic conditions were more likely to participate[30]. Bhatia et al. (2021) highlighted that chronic comorbidities can act as barriers to periodic, guideline-recommended CRC testing, indicating the need to explore cancer prevention gaps in these populations[31]. This discrepancy could be attributed to differences in the implementation of screening programs, where individuals with chronic diseases may have easier access to preventive services.

Lifestyle:

In our study, individuals with high-risk lifestyles were more likely to participate in CRC screening compared to those with healthier lifestyles. Knudsen et al. (2022) observed a trend of inconsistent participation among individuals with lower healthy lifestyle scores[32]. Carey and El-Zaemey (2019) found that in Australia, CRC screening participation was higher among those practicing healthier behaviors, suggesting that lifestyle patterns may play a key role in screening decisions[33]. Mertens et al. (2024) emphasized that identifying high-risk groups based on lifestyle data can help promote participation in both lifestyle changes and screening programs, thereby reducing the CRC burden. They also advocated for incorporating lifestyle predictors into public health models to enhance screening uptake[34].

Two studies employing Anderson’s model in the context of colorectal cancer (CRC) screening are those by Jin et al. (2019) and Chan et al. (2022). In the study by Jin et al. (2019), researchers examined Korean Americans (KAs) and found low adherence to CRC screening. They identified several predisposing and enabling factors influencing participation, such as age, household income, health insurance, regular health check-ups, doctors’ recommendations, English proficiency, CRC screening knowledge, self-efficacy, and decisional balance. Notably, enabling factors like income and doctor recommendations were predictive of adherence to screening guidelines among KAs [13]. Conversely, Chan et al. (2022) explored individual and contextual factors affecting the uptake of fecal occult blood tests (FOBT) and colonoscopy in Hong Kong. They highlighted government subsidies as the most significant enabling factor for screening participation. Additionally, perceived barriers to screening emerged as a crucial predisposing factor hindering uptake. The authors emphasized the need for ongoing promotion of subsidized screening programs and the development of targeted educational materials to address these barriers[38].

While the cross-sectional design of this study limits our ability to establish causation, it provided a practical and efficient way to analyze a large volume of real-world data. Despite this limitation, the breadth of the data allowed us to identify critical trends and associations that offer valuable insights into screening participation. These findings not only contribute to the existing body of knowledge but also lay the groundwork for future studies with more rigorous designs, such as prospective cohort studies.

Limitations and recommendations for future research

This study has several limitations. The use of registered data may introduce reporting biases. The generalizability of the findings may be limited to the population registered in the SIB system, and there may be other unmeasured factors influencing participation in colorectal cancer screening. Future research should explore the underlying reasons for the observed differences in screening participation by predisposing, enabling, and need factors level. Longitudinal studies could provide deeper insights into how changes in health policies and educational campaigns impact screening rates over time.